罗伯特·恩格尔
Robert F. Engle
(Born November 10, 1942)
经济时间序列分析专家
2003年10月８日，瑞典皇家科学院在斯德哥尔摩宣布，将2003年诺贝尔经济学奖授予美国经济学家罗伯特·恩格尔和英国经济学家克莱夫·格兰杰，以表彰他们分别用“随着时间变化的易变性”和“共同趋势”两种新方法分析经济时间序列，从而给经济学研究和经济发展带来巨大影响。
罗伯特·恩格尔1942年出生于美国纽约州的中部城市锡拉丘兹。恩格尔1964年从威廉姆斯学院（Williams College）毕业，获物理学学士，1966年和1969年在康奈尔大学分获物理学硕士和经济学博士学位。他在1969～1974年间任麻省理工学院助教，然后前往圣迭亚哥加州大学，成为该校副教授，并于1977年晋升为教授，在1990～1994年期间担任经济学系主任。目前在纽约大学任教。恩格尔的研究方向主要是利率、汇率和期权的金融计量分析，他提出了ARCH、协整（cointegration）、谱分析回归等创新性统计方法。
恩格尔担任的职务和获得的奖励有：美国艺术和科学院院士、计量经济学协会顾问、金融数量研究院墨菲（Roger FMurray）奖、美国经济研究局（NBER）研究员、计量经济学协会会员、MIT研究生经济学协会杰出教学奖等。
克莱夫·格兰杰出生于英国的威尔士。1955年获得诺丁汉大学颁发的学士学位，1959年获统计学博士学位。1974年移居美国，成为圣迭亚哥加州大学经济学院教授。随后，他开创了该学院的计量经济学研究工作，并使之成为全世界最出色的计量经济学研究基地之一。克莱夫·格兰杰的研究涉及统计和经济计量学特别是时间序列分析、预测、金融、人口统计学、方法论等领域。
格兰杰曾任美国西部经济学联合会主席。他担任的职务和获得的奖励有：美国经济学联合会年度杰出资深会员奖、斯德歌尔摩经济学院荣誉博士、芬兰艺术和科学协会外籍会员、国际预测协会会员、美国艺术和科学院院士、经济计量学协会会员等。在问鼎诺贝尔经济学奖之前，他刚刚摘取了有准诺贝尔经济学奖之称的美国经济学联合会年度杰出资深会员奖。
两位经济学家大部分的职业生涯都在加利福尼亚大学圣迭戈分校度过，并且都在2003年夏天退休。他们都还是加利福尼亚大学圣迭戈分校的荣誉教授。思格尔现在为纽约大学服务。
罗伯特·恩格尔和克莱夫·格兰杰的友谊从1975年就开始了。格兰杰说，那时候他在麻省理工大学遇到恩格尔，为他的智慧所倾倒，就邀请他到加利福尼亚大学圣迭戈分校一起工作。两个人长期合作互助，但又各有所侧重。
回忆一起合作的20多年，格兰杰在加利福尼亚大学圣迭戈分校的网页上热情地说：“我们不是同一领域的竞争对手，而是思想上相互的启迪者。我们总是一起分享研究心得，我们发表的论文都渗透着对方的观点”。
主要学术贡献
罗伯特·恩格尔 罗伯特·恩格尔的研究范围很广，跨越金融计量经济学的各个领域，但令他摘取桂冠的则是久富盛名的自回归条件异方差(ARCH:Autoregressive
Conditional Hereroskedasticity)模型。
在关于ARCH的第一篇文章中，恩格尔使用了时变性的波动率模型来研究通货膨胀。然而不久以后，人们发现ARCH最重要的应用在金融领域，因为金融市场中的活动就是对不同类型的风险进行处置和定价。在实际应用中，条件方差的变化有时会直接影响被解释变量条件期望的值。例如在考虑风险与投资回报之间的关系时．由于投资者是依据当前信息而持有证券，当风险（条件方差）增大时，投资者要求的投资补偿也就大。因此，条件方差的变化也会影响收益率条件期望的变化。与其他研究者合作，恩格尔在ARCH的基础上，建立了ARCH-M模型来分析时变风险的收益补偿。投资期里收益率取决于时变性的方差和协方差，从而自身也随时间变化。
时变性的波动率有什么实际应用呢？举个例子可以说明ARCH模型在股票收益率分析中的应用。假设用标准差表示的条件波动率在某一期间围绕05%和3%之间波动。如果投资者有一个对应与标准普尔500指数的资产组合，那么明天该投资者有多少资本面临损失？假设预测标准差是05％，他的损失(99%的概率）将不会超过资产组合价值的12%。如果预测标准差是3％，相应的资本损失将高达67%。同样，在银行和其他金融机构计算资产理合的市场风险时，在险价值(VaR:ValueatRisk)也至关重要。从1996年以来，巴塞尔国际协议规定了银行在控制资本充足率时要使用在险价值。ARCH成为金融部门风险评估中不可缺少的工具。
恩格尔现在正将他的工作拓展到不同国家间资产和发展的相关性研究。有趣的是，恩格尔从不在个人投资中使用他的模型，自称是买进持有型的典型投资者。瑞典皇家科学院称，罗伯特·恩格尔不仅是研究员们学习的光辉典范，而且也是金融分析家的楷模，他不仅为研究员们提供了不可或缺的工具，还为分析家们在资产作价和投资配搭风险评估方面找到了捷径。
对收益率的建模研究一直在计量经济学中占据很重要的位置。显然对于一阶矩的刻画是比较容易的，所以人们将注意力都放在了对二阶矩的建模上，也就是对收益率波动的计量建模。
经典资本市场理论在描述股票市场收益率变化时，所采用的计量模型一般都假定收益率方差保持不变。这一模型符合金融市场中有效市场理论，运用简便，常用来预测和估算股票价格。但对金融数据的大量实证研究表明，有些假设不甚合理。一些金融时间序列常常会出现某一特征的值成群出现的现象。如对股票收益率建模，其随机搅动项往往在较大幅度波动后面伴随着较大幅度的波动，在较小波动幅度后面紧接着较小幅度的波动，这种性质称为波动率聚类(volatility
clustering)。该现象的出现源于外部冲击对股价波动的持续性影响，在收益率的分布上则表现为出尖峰厚尾（fattails）的特征。这类序列随机搅动项的无条件方差是常量，条件方差是变化的量。
为了寻求对股票市场价格波动行为更为准确的描述和分析方法，许多金融学家和计量学家尝试用不同的模型与方法处理这一问题。其中恩格尔于1982年提出的ARCH模型，被认为是最集中反映了方差变化特点而被广泛应用于金融数据时间序列分析的模型。ARCH模型是过去20年内金融计量学发展中最重大的创新。目前所有的波动率模型中ARCH类模型无论从理论研究的深度还是从实证运用的广泛性来说都是首屈一指的。
克莱夫·格兰杰 格兰杰在学界建树颇多，其著作几乎包含所有近40年来时间序列方面的重大进展。同时，他在谱分析（经济周期分析）、因果分析、长期分析、经济预测、虚假回归和协整等许多方面的研究都是开拓性的，走在了计量经济学的最前沿。
现代时间序列经济计量学的一个重要研究课题，是探索经济时间序列数的动态结构，研究它们的统计性质，理解产生这些经济数据的生成特点和性质，从而能更有效地利用经济数据构造和建立经济计量模型，用以进行经济预测，检验各种理论的可靠性和可行性。二十世纪七十年代以前计量经济学的建模方法都以经济变量平稳这一假设条件为基础。稳定过程的特点是有一个均值，且每一时刻对均值的偏离基本相同。但在实际中，许多经济指标的时间序列都是非平稳的，并不具有固定的期望值，并且呈现出明显的趋势性和周期性。格兰杰1972年首先证明，如果直接将非平稳时间序列当作平稳时间序列来进行回归分析，可能会造成伪回归，即变量间本来不存在相依关系，但回归结果却得出存在相依关系的错误结论。经济变量表现出的非平稳性使传统建模遇到了前所未有的困难。
经济理论认为，某些经济时间序列存在长期均衡关系。例如，净收入与消费、政府支出与税收、工资与价格、进口与出口、货币流量与价格水平、商品现货价格与期货价格等。一般说来上述经济时间序列属于非平稳序列，其方差与时间成正比。看起来这些经济变量之间似乎不会存在任何均衡关系，但事实上若干个非平稳经济时间序列的某种线性组合却有可能是平稳序列。格兰杰敏锐地注意到了这一现象，利用其扎实的数学和计量经济学功底提出了协整的概念及其方法。所谓协整，是指多个非平稳经济变量的某种线性组合是平稳的。目前，协整分析己成为处理非平稳金融、经济变量相依关系的行之有效的方法。
协整理论主要用来探测变量间是否真的存在均衡相依关系，对于用非平稳变量建立经济计量模型，以及检验这些变量之间的长期均衡关系非常重要。首先，如果多个非平稳变量具有协整性，则这些变量可以合成一个平稳的时间序列．这个平稳的时间序列可用来描述原变量间的均衡关系。只要均衡关系存在，原变量间的平稳的线性组合就存在。其次，当且仅当若干个非平稳变量具有协整性时，由这些变量建立的回归模型才有意义。所以，协整性检验也是区别真实回归和伪回归的有效方法。最后，具有协整关系的非平稳变量可以用来建立误差修正模型。由于误差修正模型把长期关系和短期动态特征结合在一个模型中，因此既可以解决传统计量经济模型忽视伪回归的问题，又可以克服建立差分模型忽视水平变量信息的弱点。
格兰杰在协整概念的基础上进一步提出了著名的格兰杰协整定理，目的在于解决协整与误差修正模型之间的关系问题。该定理的重要意义就在于其证明了协整概念与误差修正模型的必然联系。若非平稳变量之间存在协整关系，则必然可以建立误差修正模型：若用非平稳变量可以建立误差修正模型，则该变量之间必然存在协整关系。在随后的工作中，格兰杰拓展了协整分析，包括处理季节趋势序列的季节协整和处理偏离超过临界值后即向均衡调整的序列的门限协整。
格兰杰不仅在理论上建立了协整概念，在现实经济分析中也做了大量的有关实证研究，引起理论界和实际政府部门的广泛注意。格兰杰学术观的一个突出特征是，一贯注重理论的现实实用性。他一贯认为和倡导经济学应该像物理学那样重视解决实际经济问题。正因为一直坚持实用的学术观点，他才从最初学习数学，转入学习统计学，到最后定位在数学、统计、经济学相结合的计量经济学研究。格兰杰一直坚持用简洁明了的文笔表述自己的学术观点，这也在一定程度上促使他的理论更加迅速地传播开来，是他著述颇丰的重要原因。
另外，作为大学教师，格兰杰在教学上颇具匠心，贯彻了他一贯的学术研究风格——注重现实应用性。他很注重对学生的启发式与实践性教育，让学生能在实际应用中对所学的经济学知识和计量经济学方法有直接的、感性的认识。
著作点击
恩格尔的主要著作有：
《协整、因果关系和预测：格兰杰纪念文集》（与怀特合编，1999）；
《ARCH：阅读精选》（1995）；
《计量经济学手册（第四册）》（与麦克法登合编，1994）；
《长期经济关系：协整阅读材料》（与格兰杰合编，1991）等。
格兰杰的主要著作有：
《经济时间序列的谱分析》（与Hatanaka合著，1964）；
《股价的可预测性》（与Morgenstern合著，1970）；
《商品价格的投机、套利和预测》（与Labys合著，1970）；
《经济时间序列预测》（与Newbold合著），1977）；
《双线性时间序列模型导论》（与Andersen合著，1978，中文版，1983）；
《经济序列建模：经济计量方法阅读材料》（1990）；
《经济学的实证建模：设定和估计》，（1999）等。
Robert F. Engle III – Autobiography
"Wake up, Robin", a gentle hand shook my shoulder. "Let's go." It was so very
early but I loved those mornings when my father would wake me and take me on
an adventure. When I was a boy and we were camping, we would leave in the quiet
morning hours to enjoy the lake together, and attempt to catch some fish. He
was an experienced fisherman and a kindly teacher. While he was in graduate
school at Cornell and the nation was still in the aftermath of the Great Depression,
he traded fish and hunted meat for room and board. I did love those mornings
and their crisp dawns.
My mother says that my father truly enjoyed having a son. My two years younger
twin sisters felt that he didn't quite know how to enjoy them. But, I wasn't
aware of those things then. So many of my childhood memories involve him. All
the excursions into science were shaped by his knowledge and enthusiasm. As
a Ph.D. chemist at Dupont, he also had access to many materials. So when my
friend, Peter Hotz, and I decided to build and shoot off rockets, my Dad supplied
different chemicals and a notebook. With each rocket we built, he would help
us construct an explosive recipe. We had to write down the ingredients measuring
carefully. He taught us to vary one ingredient at a time and then measure the
effects. We recorded everything neatly. By the time he was finished, we had
learned the scientific method and we had learned the math of measuring the distance
that each rocket traveled. Of course, we also varied the rocket design. It was
the way science should be done, passionately and carefully.
There were some experiments that Peter and I didn't tell him about. Like the
time that we built mazes and caught flies to go through them. We found it difficult
to measure their progress while they were flying, so we varied the experiment
by removing their wings. It's a little painful to think about those boyish decisions
now.
He also had a wonderfully equipped workshop where he built furniture and fixed
all manner of things. One of my favorite tools was the lathe. My mother said
that she always appreciated my constructions! As time progressed, the combination
of the science explorations and the workshop led to many science fair entries
culminating in the project that won the Philadelphia Science Fair in my Senior
year of high school. For that experiment, I built a Van de Graff generator and
ran experiments with it on X-ray transmission rates for thin metal sheets. The
generator had a big copper toilet float to collect the static electricity and
the X-ray tube was an old medical discard. A Geiger counter with homemade electronics
recorded each X-ray photon as it passed through the test sample. It was pretty
exciting stuff.
My mother was the unsung hero. She loved parenting the three of us and was always
thinking up creative projects. We still watch the home slide shows she made
of us and the neighborhood kids acting out fairy tales. She made the costumes,
wrote the plays, and drove us to scenic locations around Philadelphia (the haunted
house, the secret garden, etc). We all loved them. Except when I was supposed
to kiss my sister, Sleeping Beauty, to awake her from the long sleep. I faked
that part.
She also supervised our homework and made us feel that doing well in school
was a wonderful thing. When I was in junior high school, she arranged for me
to go to the Science Library at her alma mater, Swarthmore College. It was a
small, beautiful place with wood shelves and books to the ceiling. I relished
the time there looking through many books dreaming of when I would be able to
really master science and be in the "big time".
Both of my parents nurtured my dreams and me. It was an idyllic childhood in
many ways. We lived in a large somewhat ancient three-story house on 15 acres
in Media, Pennsylvania just outside of Philadelphia. Both of my parents grew
up in Philadelphia. My father, Robert Fry Engle, Jr., was named after his father
who owned a lovely large hotel on the Jersey shore called the Engleside. Summers
were spent in Beach Haven and school years in Philadelphia. My Dad didn't really
like the hotel business and told stories about eating in the kitchen with the
staff so that he didn't have to fuss every night with fancy clothes and fancy
food. He loved to sail and won many races. We still have silver bowls and dishes
from these events. Instead of the hotel business he headed off to Cornell where
he earned a Ph.D. in chemistry.
The Engles were a Quaker (Society of Friends) family who emigrated from Cambridge,
England in the 1600's. We assume that they came to escape religious persecution
and set up their lives in Pennsylvania with the other Quakers. My great-grandfather,
Robert Barkley Engle built the Engleside in 1876. A wonderful history of this
has been recounted by John Bailey Lloyd in his book Eighteen Miles of History
on Long Beach Island. The hotel was very successful for a time but two things
eventually led to its demise. One was the spread of the use of the automobile.
Families use to pack up and go to the beach for a month at a time bringing their
servants with them. Once the bridges were built and cars became more common
(and servants less common), families took shorter vacations bringing fewer people
along. The second factor had to do with my grandmother, Sarah Atkinson. As a
strict Quaker, she didn't believe in drinking and insisted that the hotel be
dry. This opened up an opportunity for nearby hotels who then benefited from
the additional revenue.
My father proposed to my mother on one of the round towers of the hotel. My
mother said that she thought she was marrying into a rich family with a wonderful
hotel. The hotel went bankrupt and was torn down a mere 4 years later. My father
never seemed to regret the demise of the hotel. Probably he was relieved to
know that he could proceed with his science and not be called in to take over
a business that did not interest him.
Recently, my wife, Marianne and I went to find the location of the Engleside.
There is a park and another small hotel (also called the Engleside) on the land.
Many of the old-timers recounted stories of the wonderful days of the Engleside.
It gave us great pleasure. My mother also told me that she and my father took
me to the Engleside when I was a baby just before it closed.
My mother's family had American, French, and Welsh origins. Her father, William
Vernon Phillips, immigrated to Philadelphia as a young man with his three brothers
and two sisters. They came from Cardiff and worked in the scrap metal business.
Vernon had great success. His import and export iron and steel business, F.R.
Phillips and Sons was based in Philadelphia and Milan, Italy. He was also President
of the Phillips-Laffitte Company and Chairman of the Board of the Perry Buston
Doane Company. He served as mayor of his town, president of his country club,
and deacon in his church. During World War I he was Chief of the War Industry
Board in Washington, D.C. and on the Council of National Defense. After the
Great War, he received the Knight of the Crown of Italy from the King. His obituary
called him "an eminent statesman and capitalist." He was clearly a busy man!
He married Florence Starr, a native of Philadelphia in 1912. They had three
children, Billy, Isabel, and Mary whom they nicknamed Murry. In tragic succession,
Isabel died of then unknown childhood ailments when she was 7, Billy was killed
in an automobile accident when he was 16, and Vernon died of a heart attack
in 1931 after the Crash of 1929. Murry and Florence were left on their own.
They had to move from their beautiful home to smaller quarters. Murry remembers
the sadness and loneliness of this time. One of her favorite memories is of
the European tour that she took with her mother when she was 18. The pictures
of that trip show a tall, slim beautifully dressed young woman full of charm
and grace. She attended Swarthmore College majoring in French and then took
numerous postgraduate courses in education at the University of Pennsylvania.
Although an Episcopalian, she attended Friends schools throughout her life.
When she married Bob Engle, she decided to join him in the Society of Friends
and has been active her whole life in the Quaker community.
Once her children were older, she started to teach French at Media Friends School
and eventually became its head. She sheparded the school through the addition
of a junior and senior high school and a building campaign. She remained on
its board for many years contributing both expertise and funds for new buildings
and programs.
After she and my father married in 1939, they moved to Syracuse. I was born
there on November 10, 1942 and soon after we moved back to the Philadelphia
area in Swarthmore. These were happy times. I was named Robert Fry Engle III
after my father and grandfather. Small Fry for short! They also had a favorite
Springer spaniel named Dukie. My father had taken up ice dancing and there are
pictures of my parents skating on ponds and at the Philadelphia Skating Club
and Humane Society. They would hook up a record player to a car battery and
play music while they skated outdoors.
Rob with no teeth.
Two years later, in December 1944, my twin sisters, Patricia Lee (Patty) and
Sally Starr were born. It was a shock to have three children. Grandmother Florence
Starr (nicknamed Twinkle) helped by taking me over to her apartment where I
had a wonderful time visiting the giant steam engine trains and learning card
games. In 1947, our family moved a few miles away to a big, old house in Media
that the Engles called home until my mother sold the house following my father's
death in 1983.
I was a very active child and adolescent. I loved science, sports, music, and
friends. Our yard was so big that we crafted a sloping baseball diamond. After
school, my friends would bicycle over and we would play until it got dark or
our mothers' called. Eventually in high school I learned to play lacrosse and
played goalie on the high school team. I began to play the tuba in junior high
school. My teacher said that any tuba player should also know the string bass!
I loved the instrument, took private lessons from Fred Maresch, a bass player
of the Philadelphia Orchestra and was named to the All State Orchestra. Later,
I played in the symphonies at Williams College, Cornell University, and MIT.
Eventually my bass was stolen and I took up the cello. I still have my cello
but it doesn't get much exercise.
I graduated as Valedictorian of Penncrest High School in 1960. Ours was the
second graduating class of the newly built school. There are some things that
I said in my valedictory address that I still believe. I talked about the need
for science to be both relevant and sensitive to the needs of humanity and I
emphasized the importance of a balanced life. I am still happiest when I can
do research that others can apply widely and when my days have some time for
ice dancing or skiing, cultural activities, traveling, and being with my wife
and children.
My four years at Williams College were filled with fun, growing up, and more
science. I decided to call myself "Rob" as it was more masculine and mature.
I still use this today. I majored in Physics, joined the Beta Theta Pi fraternity,
and played more lacrosse. In my sophomore year, I was named to the All American
Team as a lacrosse goalie, which was quite a thrill. By senior year I was nominated
to Phi Beta Kappa. Another delight was roaming all over the countryside going
to the women's colleges looking for dates and parties on the weekends! I seemed
to have one close girlfriend each year. Some of my best friends were made in
the fraternity house although tragically, two of them, Stan Allen and Dave Kershaw,
died young. They would have hooted with pleasure over the Nobel honor.
Academically, I started Physics and Math at the sophomore level since I had
done advanced work and lots of outside reading in high school. A small event
happened in my senior year that would have great consequences. I could take
one more elective to finish out my courses. I almost took the religion course
that everyone said was great, but I decided instead to try Introductory Economics.
My roommate, Walt Nicholson, and many of my fraternity brothers were economics
majors and spent hours discussing economic issues. I loved the course. It was
interesting and came easily to me. Who could have predicted that this one elective
would later serve as the catalyst that changed the course of my life?
During my senior year, my advisors in Physics encouraged me to apply to graduate
school. It seemed natural since I had always dreamed of being a scientist. But
something was lacking. I was losing my passion for physics. I did finally apply
at the last minute and was accepted at Cornell and UC Berkeley. I called to
accept Berkeley, but it was lunchtime. By the time they called back, my advisors
had suggested I go to the famous labs at Cornell and so I did, following in
my father's footsteps. I graduated from Williams Cum Laude with Highest Honors
in Physics and went off to Ithaca, New York.
Graduate school had its own pleasures. I lived initially in Sage Hall, which
was the graduate student dorm. It was wonderful to be on campus with both men
and women. I began to date a girl who was an ice skater who took me to the Cornell
Figure Skating Club to learn ice dancing. I was very pleased to be in an environment
with families and skaters of all ages. As a child my father tried to get me
to skate with him, but I had no interest. Now, I needed a new sport since my
lacrosse days were over. Ice-skating began to look like a lot of fun. Men are
always in short supply in ice dancing so there were plenty of talented pretty
girls to help teach me. There was also Eddie Collins, a former Canadian competitor,
who took me on as student. The girl friend quit the club and me, but I have
enjoyed ice dancing since then.
I worked in Professor Watt Webb's lab doing low temperature physics and studied
quantum mechanics from Nobel Laureate Hans Bethe. Watt Webb was a great advisor
with lots of insight and creativity and the study of super-conductivity was
certainly exciting. However by mid-year, I realized that I did not want to spend
the rest of my life as a physicist. Physics wasn't what I expected, or maybe,
what I expected was what I realized I did not want anymore. Working in the bowels
of a building on projects that interested very few people in the world held
less interest for me. After agonizing over these feelings and talking to my
friends, I approached Alfred Kahn, chairman of the Economics department at Cornell.
I asked if his department would be interested in considering me for their Ph.D.
program. Serendipitously, he said that an NDEA fellowship had just become available
because the intended recipient had turned it down. He offered it to me but said
that I needed to make an immediate decision or they would offer it to someone
else. My head spinning, I accepted the offer. Now, I had to tell my father and
Professor Webb. Both were supportive but with heavy hearts. My father had trained
me to be a scientist and now I was leaving for one of those "soft" science fields.
At that point, neither he nor I realized how important all that scientific training
would be in my contributions to economics.
I loved economics. I attacked it with an energy that I had lost. I took undergraduate
classes to make up my deficits in knowledge while I finished my Masters in Physics
on using nuclear magnetic resonance to study the performance of a high temperature-superconducting
magnet. Intriguingly, both superconductivity and NMR, now named MRI, were awarded
Nobel prizes this year! In fact the superconductivity lab next to Webb's had
been honored by the prize in 1996 and they were in the audience this year.
As I began the graduate program, my new advisor, Ta Chung Liu, was in Taiwan
working with Chiang Kai Shek on their economic development plan. I took courses
in Econometrics from Berndt Stigum where we read Malinvaud in its new English
edition, and with John Fei who introduced me to modern microeconomics. My knowledge
of math and statistics was put to use instantly. I took Kiefer's probability
and Wolfowitz's statistics. I was extremely happy.
The next year, Ta Chung was back and I took his Econometrics class which gave
me a solid basis for understanding the field. He was working on his third model
of the U.S. economy. He had built an annual model and a quarterly model, having
beaten the enormous Brookings Project to this goal. Now he was building a monthly
model with all the new data and econometric problems that engendered. I learned
trade from Ron Jones and Jaroslav Vanek, growth and development from John Fei
and economic history from Woody Fleisig. It was so much fun. I found that I
could solve the problems easily but quickly got stuck if I needed to formulate
a problem. The casual discussions of short run and long run elasticities always
required tremendous thought. It was probably ten years before I really absorbed
the economic way of thinking.
The years spent in graduate school passed quickly. A few summers were spent
in Washington, D.C. at the Bureau of the Budget doing program budgeting. I worked
for John Deutsch who later became CIA director. It was there that I first tried
to solve a cost benefit problem for public transportation. I recognized that
land use was endogenous in the long run and this was critical for the analysis.
This was my introduction to urban economics. I met John Kain and Charles Schultz.
As the years passed, my work became more focused on time series econometrics.
Ta Chung suggested that I try to theoretically analyze the relationship between
the different time scales for economic modeling. I was able to use some of my
physics skills by formulating the problem in the frequency domain and applying
Clive Granger's "typical spectral shape" for an economic time series. This was
my first introduction to his work.
In January of 1968 another important event occurred. After a ski trip to Aspen,
Colorado with friends, I returned to start the second semester. There was a
party down the street and a beautiful new young woman caught my eye. She had
graduated early from a California college and was just beginning graduate study
in Child Development at Cornell. Her name was Marianne Eger and she became my
wife in August 1969.
We were first good friends and then I asked her to come with me and my friend
Linus Schrage on a week-long canoe trip to the Algonquin Wilderness in Canada.
She had never slept in a tent or been in a canoe but she thought it might be
fun. Her sense of adventure has been a hallmark of our married life. We love
traveling the world, hiking, skiing, and snorkeling while meeting new people
and appreciating natural environments. We have come to understand many of the
world views that exist on this planet. Our children have embraced the sense
of adventure and we continue to have great fun together.
She is also a wonderful intellectual companion and mother. Although I do not
understand the intricacies of psychological theory and she, likewise, is not
a trained economist, we discuss our ideas and careers with each other at length.
And our children, Lindsey and Jordan, are the joys of our lives. I consider
myself a very fortunate man.
Marianne was born in Presov, Slovakia and immigrated to the USA in 1949 when
she was 2. Her mother, Edith Elefant, grew up in Kosice and was a teenage survivor
of Auschwitz. Her parents perished there. She met her husband, Albert Bela Eger,
at a recuperation spa after the war. He was a resistance fighter from Presov.
After WW II, the Communists made life uncomfortable and they left after an assassination
attempt. They were fortunate to receive a visa and joined his brother and her
sister in Baltimore, Maryland. They had to leave the Eger fortune behind and
like many immigrants, found a way to reinvent their lives. Bela became an accountant,
Edith, a bookkeeper. Marianne, meanwhile, went to nursery school. The family
moved to El Paso, Texas to pursue more lucrative opportunities. El Paso was
good to the family but Marianne was ready to leave. She went to Whittier College
on scholarship, graduated early and then went on to a Cornell.
Although our marriage made finishing her Ph.D. more difficult, Marianne did
receive it 1978, two years after our daughter Lindsey, was born. She has gone
on to have a very successful career as a clinical child psychologist and as
a sports psychologist. She has an active speaking career. She even wrote a weekly
food column for a number of years.
Rob, Marianne, Jordan and Lindsey in Williamstown 2002.
But I've gotten ahead of my story. We married on August 10, 1969. On that day,
I turned in my dissertation, received my Ph.D. and we left Cornell for good
to take my first academic job at MIT. It was strange to arrive at a place with
so many famous economists, none of whom were particularly interested in time
series. For one overlapping year, Christopher Sims was at Harvard and a few
years later, Jerry Hausman came to MIT. They were both helpful and enjoyable
to talk with.
The first summer, 1970, I developed the theory for Band Spectrum Regression
and attended the World Congress of the Econometric Society in Cambridge, England.
I met economists there who have become lifelong friends and collaborators -
Clive Granger, Ken Wallis, David Hendry. I began to go to London and the LSE
every chance I could to pursue my fascination with time series. But finding
my place was still complicated at MIT.
I enjoyed the teaching and the students there. Many of my students from that
time have gone on to do quite well themselves: Larry Summers, Larry Backow,
Ric Mishkin, Hal White, and many others. I even received an Outstanding Teacher
of the Year Award from the graduate students. But I thought that my research
needed a new focus.
Frank Fisher, Bob Solow, and Jerry Rothenberg encouraged me to join them on
a new project to build a model of the city of Boston. A fellow economist and
friend from Cornell, Tom O'Brien, was research head of the Boston Redevelopment
Authority. Over the next 5 years and with many graduate students, the model
took shape. I became an urban economist publishing very elaborate statistical
models in a field not known for its mathematical sophistication.
Although MIT promoted me to Associate Professor, it was clear that I would not
get tenure there. I wanted to work more in time series and I was attracted to
Clive Granger's interest in spectral analysis. At a meeting in Washington, D.C.,
I asked him if his new university, UCSD, had any openings. He said he'd check
and I was invited to La Jolla, California to give a talk there. It was mid-winter,
they housed me in a lovely hotel on the beach, a Williams Beta, Dick Attiyeh,
was Economics Department Chair, Clive Granger had recently arrived - it was
irresistible. I was hired as an Urban Economist and for years I did teach their
urban economics course.
It was the beginning of a golden time for time series econometrics. Clive and
I began an econometric seminar series, hired Hal White and later Jim Hamilton.
We had funding for visitors and the best and brightest came from all over the
world. We had numerous visits from David Hendry, Svend Hylleberg, Søren
Johannson, Katerina Juselius, Timo Teräsvirta, Ken Wallis, Grayham
Mizon, Tony Hall, Adrian Pagan, Max King, Giampierro Gallo, Tony Espasa, Keith
MacLaren, Bernt Stigum, Eilev Jansen, Øivind Eitrheim, Helmut Lütkepohl,
Hermann van Dijk and many others.
I took a sabbatical at LSE in 1979. Lindsey was then 2 and we rented a lovely
row house in Hampstead with a back garden and a study in the back of the first
floor. It was there that a new idea came.
I was interested in Milton Friedman's conjecture that inflation uncertainty
was a central cause of business cycles. Investors who did not know what prices
and wages would be in the future might invest less. To test this, a time series
model was needed with variances that could change over time. There were two
tools that came together to solve this problem. I had done a lot of work with
the Kalman Filter and recognized that a one step predictive density would be
sufficient to define a likelihood function. The second tool was a test. Clive
had recently proposed a test for bilinear time series models. He came by my
computer one day before I left and suggested I square the residuals and then
fit an autoregression. To my amazement, it was quite significant. I suspected
that this test was the optimal Lagrange Multiplier test for some new type of
model, but not the bilinear model. I was later to discover that it is indeed
the optimal test for ARCH and it is so called today.
Lunch and tea at the LSE were very stimulating times for me. Each day I would
get a little further on this new model and would talk with Sargan or Durbin
or Hendry or Harvey about its properties and my proofs. David Hendry eventually
named it AutoRegressive Conditional Heteroskedasticity and offered to have Frank
Srba program it. We applied it to UK inflation data and the ARCH model was launched.
Looking back now, one might think that new ideas are easy to publish. At least
for me, they are not. It took quite a bit of rewriting and persuading to finally
get it accepted in Econometrica. In fact, I don't think that any of my papers
have had an easy time of it!
Coming back to La Jolla, I was pleased that two events had occurred. The ARCH
model was on its way and so was our son, Jordan, who was born in May 1980. I
became very interested in Clive's new concept of low frequency correlation that
he called cointegration. It seemed to me that the new tests for unit roots of
Dickey and Fuller, which had been so successfully applied by Nelson and Plosser,
could be extended to this case. So I constructed an econometric approach to
estimation and testing of cointegrated systems. Initially we wrote two joint
papers and presented these at a time series meeting at UC Davis. There was a
lot of discussion with many opinions as to whether this was a big or a small
innovation. We decided to put these together but again, it took rewriting and
persuasion, before Econometrica accepted this one too. Clive lost patience and
published another version of the paper in the Oxford Bulletin.
The ARCH Cafe in Sydney. My favorite!
Also, my excellent students, Mark Watson and Tim Bollerslev were busy. Mark
had carried the Kalman Filter and state space models to a wide range of problems
in macroeconometrics. We wrote a paper combining factor analysis and time series
modeling and called it DYMIMIC. His first job was at Harvard where he worked
with Jim Stock on models of leading indicators. This eventually became an official
NBER business cycle forecasting model. Tim took the ARCH model, added a moving
average and created GARCH. The GARCH model is an infinite order ARCH model with
a geometrically declining set of weights. This extension made the model even
more useful. The simplest GARCH model often performs successfully in a wide
range of data.
Although these lines of research are now the most visible, there were a lot
of other interesting things going on. I had visited LSE in 1975 on a Ken Wallis
grant. There I learned about Lagrange Multiplier Tests and wrote a series of
papers of which the Handbook of Econometrics survey is the one that gets the
most attention. I gave a talk at C.O.R.E. and got into a big argument with Jean
Francois Richard on the meaning of exogeneity. Afterward we looked up Koopmans'
definition and realized there were two different notions which we dubbed weak
and strong exogeneity. The difference between these two was Clive's concept
of "Granger Causality" which was being used as a test for exogeneity by many
authors. We invited David Hendry to join us, added a characterization of "super
exogeneity" to deal with the "Lucas Critique" and wrote a quite controversial
paper. After the dust settled, the paper was published by Econometrica and the
profession has more or less accepted these exogeneity concepts.
Ramu Ramanathan, Clive and I founded a small consulting company, QUERI, that
was dedicated to doing econometric research. We spent many happy hours on EPRI
(Electric Power Research, Inc.) and other projects working with excellent graduate
students to estimate better electricity demand systems, or load factor systems,
or weather sensitivity models. These models have found an important place in
electric utilities and in EPRI research, but only the spline model done with
John Rice has received much academic attention. I wonder whether we would have
been involved in the California energy crisis if QUERI had still been in existence?
I loved going to the European meetings of the Econometric Society. Time series
was a big part of the program each summer. The meetings were invariably in nice
places and we developed long term friendships that were renewed annually. By
the mid 80's these meetings featured multiple sessions on cointegration and
ARCH. It was so exciting to see the new directions people were taking this research.
We kept working to keep up with these fast flowing fields of research. Each
year we would have several students working on these topics. Many of these students
have had highly successful careers in and out of universities.
The profession rapidly developed a wide range of extensions of ARCH and GARCH
models. Dan Nelson who sadly passed away at a very early age, introduced the
Exponential GARCH model called the EGARCH. This model allowed an asymmetric
response to returns. This was followed by many versions with names - TARCH,
GJR-GARCH, NARCH, VGARCH, APARCH, FIGARCH, FIE-GARCH, STARCH, SWARCH, CES-GARCH,
SQGARCH, component GARCH and many more. It was an exciting time. Students such
as Gary Lee, Ding, Raul Susmel, Victor Ng, Ted Hong, and Ray Chou contributed
to this growing literature.
As my research career developed, so did my ice skating hobby. I began skating
with Brigit Luciani and we competed in many adult competitions. We often were
successful in regional competitions. I passed my gold dance test around 1990
and shortly after that we did a week long workshop with Jane Torvill and Christopher
Dean, legendary innovative ice dance champions. Finally, the United States Figure
Skating Association agreed to have a national adult skating competition. The
first year was 1995 in Wilmington Delaware. There were 21 teams and we eventually
ended up 4th. We were very pleased. In 1996 we managed to move up to second
place. Later I began skating with Dr. Wendy Buchi and we were again quite successful,
placing second in 1999. I've had lots of incredible coaches over the years including
Barret Brown, Michelle Ford, Jeannie Miley, Kent Weigle, Tatiana Navka, Judy
Blumberg, and now Natalia Dubova and Eve Chalom. When I am skating, economics
is far away. I always return refreshed and ready to carry on.
Ultimately, my interests gravitated more and more to finance. My colleagues
David Lilien and Mike Rothschild and wonderful graduate students, Russ Robins,
Victor Ng, Ken Kroner, Mustafa Choudhury and Aaron Smith helped me see this
new way of thinking. The trade-off between risk and return was a central feature
of financial analysis and the ARCH model had a mechanism for measuring this.
We called it the ARCH-M model. A multivariate version with Tim Bollerslev and
Jeff Wooldridge generalized the Capital Asset Pricing Model, CAPM. Options markets
traded volatility and gave a quantitative validation of the ARCH models. Alex
Kane, Jaesun Noh and I explored the possibility of using GARCH to trade options;
it was temporarily successful. We flirted with the idea of managing money or
investing our own but ultimately didn't take that route. I ran several conferences
in San Diego on the ARCH model designed for both academics and practitioners.
I remember being so delighted that finance faculty such as Michael Brennan,
Eduardo Schwartz and Bill Schwert came. It was a peek into a new profession.
I knew Michael from MIT where I had stepped in to be part of his thesis committee
when I first arrived. He was about my age and we still joke about how he is
my first student. The rest of the committee was Franco Modigliani and Myron
Scholes - a hat trick!
David Hendry and Rob after a great meal at the Waterside Inn outside London.
Gradually this new area of research became known as financial econometrics.
It is the development of statistical tools specifically designed for financial
applications. At conferences and meetings there were now quite a few sessions
on financial econometrics. At UCSD we developed a specialization in financial
econometrics. The arrival of Bruce Lehman and Allan Timmerman were very important
in buttressing the finance side of this research area.
As these ideas spread around the globe, I was increasingly asked to give long
workshops or mini-courses. This was fun as a travel experience and introduced
me to students I still meet. The first was in Nairobi where Marianne and I were
able to spend time with Jim Tobin and his wife. We did a short course at Peoples
University in Beijing in 1995 flanked by Gregory Chow and Angus Deaton. Mini-courses
in Maastrict, Bagni di Lucca, Helsinki, Uppsala, Madrid, London, Sydney, Stockholm,
Vienna, Vaasa, and Rotterdam followed and gave opportunities for me to survey
research in particular areas.
Joint work with Sharon Kozicki, Farshid Vahid, Joao Issler and Raul Susmel introduced
another idea into econometrics - common features. This is a generalization of
factor analysis that has applications to cointegrated systems, multivariate
ARCH models and many areas of economics. The goal is to simplify multivariate
systems into a small number of variables that endow the system with its typical
features. In this case the features are common. This simplifies analysis in
a way that is widely used in financial research. I will be surprised if there
aren't important research developments using common features in the next few
years.
Somewhat later Simone Manganelli and I developed a new approach to measuring
the extremes of a distribution. This method quantifies the probability of large
losses on a portfolio by predicting the future quantiles. It has the tastiest
name of any of my models, the CAViaR model. It stands for conditional autoregressive
value at risk. Peter Hansen came up with this clever mnemonic. The CAViaR model
uses the theory of regression quantiles in a time series context to give an
updating formula for the Value at Risk (extreme quantile) of a portfolio.
In the early 1990's I gave my first talks to financial practitioners. In a series
of RISK meetings and Q-Group meetings, I introduced the audience to ARCH models
and learned about the fascinating questions facing practitioners doing risk
management and derivatives trading. I began a consulting project at Salomon
Brothers in building seven of the World Trade Center. This project with Joe
Mezrich and Eric Sorensen was to incorporate the GARCH model into a series of
client oriented trading systems. Volatility forecasts and trading strategies
were proposed based on innovative models. Advanced software developed by Pat
Burns and Aslihan Salih, was used to structure client and proprietary portfolio
strategies. I learned a great deal from this experience and made many contacts
throughout the financial world.
I was asked to serve on the steering committee of a Zurich Financial Services
company, Olsen and Associates, who were doing path breaking work on the analysis
of very high frequency financial data. They were partly data vendors who delivered
tick by tick currency and other data to a collection of European clients, and
partly model builders. Richard Olsen and Michel Dacarogna proposed joint academic/practitioner
conferences on the use of these data. In the first high frequency data conference,
the same data set was analyzed by statisticians, economists, physicists, financial
practitioners, and traders. The comparison was fascinating and highly informative.
Jeff Russell and I wrestled with the question of what we could learn from such
high frequency data. Every way of handling the data in calendar time involved
a loss of information. One day we realized that we could treat the time between
trades as a random variable itself and model the speed of trading. We developed
the Autoregressive Conditional Duration or ACD model for this purpose. It is
a Poisson process with an intensity that is conditional on the past information.
Financial data show variations in trading rates that can be called duration
clustering. The model we proposed had the same structure as the GARCH model
and recognizes the close relation between trading intensity and volatility.
The simple ACD(1,1) very rapidly led to generalizations with different functional
forms, error densities and types of data. Initially we only observed quotes
so we built a model for the time it took for prices to move a fixed amount.
The ACD model of such price durations is really an alternative volatility model.
Then we obtained transactions data and modeled the time between trades. In the
Fisher Schultz Lecture that I gave in Istanbul, I combined the time between
trades and the prices at which trades occurred into an "ultra high frequency"
or tick by tick GARCH model.
Although there was a lot of interest in the statistical model of trade arrivals,
some economists at the time thought this was a model of the least interesting
variable in finance. I was convinced that the trading frequency measured a fundamental
heartbeat of financial markets. Clearly it reflected the flow of information.
It turns out also to be closely related to measures of liquidity. I began reading
the literature on market microstructure and soon came to Maureen O'Hara and
David Easley (1992). Maureen was also on the Olsen steering committee and I
recognized that the speed of trading was related to information flow in their
model and had a direct link with measures of liquidity such as bid-ask spread
and price impact. In papers with Alfonso Dufour, Joe Lange and Andrew Patton,
we showed that when markets are more active, in the sense that the time between
trades is short, they are less liquid. Information flows correspond to illiquid
markets over time.
This research leads naturally to an interest in timing of trades to achieve
good execution. Several years later, a group at Morgan Stanley headed by Robert
Ferstenberg and Rohit d'Soza approached me and I have spent many interesting
days with them developing a microstructure approach to optimizing and evaluating
trades. We have become good friends as we push this frontier between theory
and practice.
New York was a very exciting place to visit when I was consulting for Salomon.
Now that Lindsey and Jordan were both on their own, I convinced Marianne to
spend the fall semester of 1999 visiting NYU in the finance department. I knew
many people there. My sister, Patty, had also moved there to work as the child
development officer for UNICEF. Joel Hasbrouck was the leading microstructure
econometrician; Steve Figlewski, who was a student of mine many years ago at
MIT, was a leading empirical options researcher, and my recent student Josh
Rosenberg was there. The semester was too short. I saw the financial markets
at work and got to know more finance faculty. We ate well and played hard. When
NYU offered me a permanent position, I took it and we moved in September 2000.
I thought we would stay for only a year, but even that was not enough. Eventually
I retired from UCSD, becoming an emeritus professor. We maintain close ties
with San Diego as we have a house there where we spend summers and have lots
of friends. UCSD has kept my office and I still see a few students.
Since I have been in New York, my work has actually moved back to volatility
models, but in large multivariate systems. The extension of ARCH models from
univariate processes to multivariate processes began in the early 1980's. Ken
Kroner and I introduced one new family, he and Victor Ng introduced another
and Tim Bollerslev introduced a third. However, there have not yet been many
empirical applications of these models. This is because they are difficult to
specify, estimate and interpret. In a new general class of models called Dynamic
Conditional Correlation or DCC models, I have developed a potential solution.
This class is parsimonious and appears to give satisfactory performance for
both small and large systems. In a series of lectures at Erasmus University
in May 2003, this model was fully described and these lectures will soon appear
as a Princeton University Press monograph.
To evaluate alternative covariance matrices, I have developed a loss function
with Riccardo Colacito, based on the effectiveness for asset allocation. The
same model can be applied to Credit Risk as the correlation between defaults
is determined by the dynamic structure of the covariance matrix and the tail
properties of the model. New high frequency volatility models have come out
of research with Giampiero Gallo and models for the volatility of volatility
have been developed with Isao Ishida. The new and interesting research topics
that have opened up since I have been in New York are a constant source of stimulation.
While professional interests have always been important in my life, so have
my family and my hobbies. Marianne and I have had a wonderful busy time raising
Lindsey and Jordan. We stay as close as we can to them and over the years they
have traveled with us all over the world. Lindsey was always a fine student
and devoted ballerina. Her warm, friendly personality and her strong determination
have led to much success for her and a good time for us. She has been a wonderful
sister to her younger brother. When she was a baby, she would come to my office
one day a week and sleep on a mat behind a chair to take her nap. She would
come to lunch with Clive and the other economists. Unfortunately, all that good
input didn't lead to a career in economics! She graduated from Princeton Cum
Laude in anthropology and has now finished a Ph.D. in developmental psychology
at UCLA. In May, 2003, she married Justin Richland, JD, a legal anthropologist
from Los Angeles, also finishing his Ph.D. They both hope to have careers in
academia.
Jordan is the outgoing charmer. He is a wonderful athlete, excellent soccer
player, and has a probing mind. He graduated from Williams College in English
literature with a certificate in theater and is now in Los Angeles working as
an actor and cinematographer. During his college years, he spent 2 semesters
at the University of Cape Town, South Africa. He was very affected by the country
and its people. Our trip with him to Botswana was a highlight for us. Jordan
can talk to anyone comfortably and learn about them. He is also an excellent
writer. His career path will be an interesting one. As a brother and a son,
he is a pleasure for all of us. He has had lots of girlfriends and was especially
sorry that the Swedish princesses did not seem interested in him!!
My sisters Sally and Patty have also had very successful careers and I am so
proud of them. It is remarkable how many parallels there are among us. I guess
it must be something about either genetics or our upbringing. Sally received
her Ph.D. in anthropology and has been teaching at Wellesley for 30 years. She
developed an interest in the relation between legal systems and cultural systems.
This lead to books on Urban Danger in Boston, the ways of Getting Justice and
Getting Even in and outside the Massachusetts court system, law, culture and
the U.S. colonization in Hawaii, human rights and the U.N. and lots of papers
with sharp insights into modern society. She was president of the Law and Society
Association and a regular visitor at the American Bar Foundation. She is now
being courted by NYU and the University of Pennsylvania. Maybe there will be
another Engle in New York.
Patty received her Ph.D. in psychology at Stanford and has spent a career studying
cross cultural child development focusing often on the role of nutrition. She
has been involved in projects in Guatemala, Nicaragua, Peru, Uganda, India,
and many other places. Her academic publications describe many of these research
settings; her students and colleagues at her long time university, California
State University at San Louis Obisbo, were able to hear about these first hand.
Her current position as Senior Advisor for Child Development at UNICEF in New
York gives her scope to pursue the goal of child health from a broad perspective.
She has become a vocal advocate for global children as she works to improve
the quality of life in so many different cultures.
Rob and Wendy Buchi in National Adult Ice Dance competion in Lake Placid 1997.
A stable, happy home life is good for all of us and allows us each to focus
on our careers knowing that when we are back together, there is more to share.
As a family we love food, wine, music, opera, art, theater, traveling, hiking,
fishing, and watching sports. Separately we have our individual career paths
and our hobbies. It makes for a rich interaction. I still ice skate 2-3 times
a week. I enjoy skating with Wendy Buchi when I am in La Jolla. Marianne is
learning to play golf. Lindsey runs with her dogs, Jordan finds soccer games
and surfs when he can.
The Nobel Prize in Economics is an incredible recognition for the work that
my students, colleagues and I have done over the years. We all worked hard but
we were also lucky that the financial applications were so important. It continues
to amaze me how far this simple idea has traveled. I am starting a Financial
Econometrics Research Center at NYU to foster the continuing development of
this field.
As I look back over my career, this Prize is the high point. I find myself reflecting
with great affection on smaller, perhaps less dramatic, moments. These are moments
of insight; moments that started a new research topic or recognized a connection
between things previously thought to be distinct. These are also moments of
family time - wonderful moments with Lindsey and Jordan, sharing ideas, eating,
hiking, skiing and yes, even fishing - and moments with my lifelong companion
and soul-mate, Marianne. It has been already a lengthy and varied career but
I think I am only part way through it. There are many exciting adventures in
our future and I am looking forward to ever so many special moments more.
From Les Prix Nobel. The Nobel Prizes 2003, Editor Tore Frängsmyr,
[Nobel Foundation], Stockholm, 2004
This autobiography/biography was written at the time of the award and later
published in the book series Les Prix Nobel/Nobel Lectures. The information
is sometimes updated with an addendum submitted by the Laureate. To cite this
document, always state the source as shown above.
【关闭窗口】